Robotics paper index
Structure-Aware Robust Fine-Tuning: Defending Vision-Language-Action Robots Against Physical Attention Hijacking
One-line summary
A robotics research paper on Structure-Aware Robust Fine-Tuning: Defending Vision-Language-Action Robots Against Physical Attention Hijacking.
Engineering notes
Engineering notes will be added by the Robot Papers editorial team.
Chinese explanation / 中文解读
中文解读待补充:本站会优先为 VLA、具身智能、人形机器人控制、机器人操作等高价值论文补充中文说明。
Original abstract
Vision-Language-Action (VLA) policies promise general robotic manipulation, but their robustness against physical-world attacks remains fragile. In particular, we show that physically realizable adversarial patches can reliably induce failures by triggering a mechanism we call policy-critical action-to-vision attention hijacking, where action-conditioned attention is diverted from task-relevant regions to a localized patch. To demonstrate the threat, we propose Attention-Guided Semantic Disruption (AGSD), an Expectation-over-Transformation (EOT) optimized printable patch that jointly (i) concentrates action-to-vision attention on the patch and (ii) disrupts vision-language semantic alignment, yielding strong cross-task and cross-architecture transfer. To mitigate such attacks, we introduce Structure-Aware Robust Fine-Tuning (SARF), a zero-inference-overhead defense that fine-tunes only the visual encoder using feature anchoring, policy-critical attention correction, and language-guided geometric consistency restricted to semantically relevant regions. On LIBERO, SARF reduces OpenVLA's failure rate under AGSD from 100% to 14.2%-56.8% (28.6% average) across suites while preserving clean performance, and on a real PiPER manipulator it improves average success under AGSD from 23.0% to 65.0%. These results highlight mechanism-level robustness as a practical path to securing VLA robots against physical attention hijacking.
Links and sources
Need this topic turned into a technical roadmap?
Robot Papers can prepare a custom robotics literature review, code map, dataset map, and B2B technology assessment.
Request B2B research
Comments